Dear All,
We are pleased to share the second call for participation for AlexandriaX-2026, an ArabicNLP 2026 Shared Task on Context-Aware Dialectal Arabic Machine Translation and MT Evaluation.
We are excited to announce that the competition platform is now live. The training/development data, baseline code, evaluation resources, and submission platform for the AlexandriaX-2026 shared task subtasks are now available for participants through the official website:
AlexandriaX-2026 aims to advance research on dialect-aware, context-sensitive, and diagnostically interpretable Arabic machine translation. The shared task focuses on realistic Arabic MT scenarios involving dialogue context, dialectal variation, speaker and contextual metadata, cross-dialect translation, and fine-grained MT error analysis.
Participants are invited to work on one or more of the following subtasks:
We warmly encourage research teams, students, NLP practitioners, and members of the Arabic NLP, machine translation, and computational linguistics communities to register, explore the released resources, and participate in AlexandriaX-2026. Participants may choose to participate in any subset of the subtasks.
Please make sure to register and join the Google Group to receive updates concerning the shared task, including announcements related to data, evaluation, submission instructions, and results.
Website and resources:
https://alexandriax.dlnlp.ai
Registration form:
https://forms.gle/K6d9B5Qs5EsJTF2V7
Google Group:
https://groups.google.com/g/alexandriax-2026
Important dates:
We look forward to your participation and to seeing the community’s contributions to more realistic, dialect-aware, context-sensitive, and interpretable Arabic machine translation.
Best regards,
The AlexandriaX-2026 Organizing Team